flyingby/Awesome-Deepfake-Generation-and-Detection

What it solves

Research on deepfakes is scattered across many papers, venues, and techniques. This project collects and organizes the most relevant work on facial manipulation and its detection into one place. It helps researchers and practitioners quickly find published methods for face swapping, face reenactment, talking face generation, face attribute editing, and forgery detection, along with related areas like head swapping, face super-resolution, and makeup transfer.

How it works

The repository is a structured, up-to-date list of academic papers. Each paper is listed with its year, publication venue, category (like GANs, diffusion, VAEs), and a link to the paper and, when available, its code. The content is grouped into main sections for the five core tasks, plus a section for related research domains. The maintainers also point to a comprehensive survey paper (published in ACM Computing Surveys 2026) that summarizes these methods, and they invite contributions via pull requests to add any missing work.

Who it’s for

Researchers, students, and engineers working on deepfake generation or detection. It is also useful for anyone who wants a structured overview of the current state of facial manipulation technologies, especially those interested in the academic literature or in building or evaluating detection systems.

Highlights

• Includes a broad survey covering face swapping, reenactment, talking face generation, attribute editing, and forgery detection.
• Provides detailed results for the most representative works, helping readers compare approaches.
• Contains many related domains, such as head swapping, face super-resolution, face reconstruction, face inpainting, body animation, portrait style transfer, makeup transfer, and adversarial sample detection.
• Lists papers from top venues (CVPR, ICCV, NeurIPS, ECCV, etc.) with direct links to code when available, making it easy to dig deeper.